claw-skill-guard

To scan a skill for malware, run the scanner script against the local directory, remote URL, or ClawHub slug. This identifies malicious patterns, suspicious URLs, and installation traps before the skill is installed.

Skill vulnerability assessment detects malicious code hidden in AI agent skills by analyzing files with predefined regex patterns and an allowlist. It categorizes identified threats into critical, high, medium, and low-risk levels to ensure code safety.

Yes, you can scan remote URLs and ClawHub skill slugs for security threats. The scanner evaluates these sources for dangerous commands and malicious patterns before you install them on your system.

You can integrate security scanning into your CI/CD pipeline using pre-commit hooks and AGENTS.md enforcement guidance. This automates threat intelligence checks to block malware during the development workflow.

Regex-based malware detection relies on predefined patterns and an allowlist, meaning it may miss novel or obfuscated threats not yet recognized. It categorizes known suspicious URLs and installation traps but requires updated patterns for new vulnerabilities.